Figure 1.
Membrane deformation by human pacsin isoforms.
A. Domain organization and structure of pacsin-1. The structure shows a F-BAR domain dimer with the protomers shown in green and orange, respectively. B. SAXS-based comparison of full-length pacsin-1 in solution and in crystals. Distant distribution functions, Rg and Dmax values were determined based on the full-length crystal structure [28] and the solution scattering data [27]. Rg/Dmax (crystal) = 212/60 Å; Rg/Dmax (SAXS) = 215/58 Å. Discrepancies between the respective distance distribution functions can be explained by the flexible linkers that connect the F-BAR and SH3 domains and were not modeled in the crystal structures. C. Negative-stain electron micrographs. The membrane deformation potential of human pacsin isoforms and their isolated F-BAR domains was monitored by EM. Folch fraction I liposomes were incubated with purified proteins (5–10 µM), and processed as described in Materials and Methods. Arrows indicate specific membrane morphologies (solid arrows, pearling structures; dashed arrows, narrow tubules; open triangles, wide tubules). Inset shows liposome-only control; scale bar, 100 nm.
Figure 2.
Activation of pacsin-1 by the proline-rich domain (PRD) of dynamin-1.
A. Sequence of the mouse dynamin-1 PRD. A regulatory sequence (phospho-box), the core pacsin-1 binding region (orange) and arginine residues are highlighted. The sequence is 100% identical to the human dynamin-1 PRD. The mouse PRD was expressed as GST-fusion protein. B. Negative-stain EM with Folch liposomes. Liposomes were imaged as described before following incubations with the indicated proteins and protein complexes (top panel). The histogram (middle panel) shows the size distribution of the vesicles produced by pacsin-1 in the presence of GST-PRD. Vesicle diameters were quantified from electron micrographs taken from three independent experiments. Liposome-protein co-pelleting assays (bottom panel) were used to assess the amount of protein bound to lipid vesicles. The horizontal, dashed line indicates the lipid-bound fraction of the isolated pacsin-1 F-BAR domain under similar conditions. Two-tailed unpaired t-tests for both pacsin-1 and GST-PRD were p<0.05, N = 4. C. Negative-stain EM with synthetic lipid mixtures. Experiments were carried out as described in (B), but using liposomes with the composition POPC/ POPE/ POPS = 27.5/27.5/45. Error bars represent standard deviations of a minimum of 3 independent experiments.
Figure 3.
Effect of GST-PRD truncation mutants on the membrane deformation activity of pacsin-1.
A. Membrane deformation of Folch liposomes. The sequences of mouse dynamin-1 PRD truncation mutants GST-PRDtrunc1 and GST-PRDtrunc2 are shown (top panel). Negative-stain EM images are shown after incubation of liposomes with the indicated protein complexes. Either wild-type human pacsin-1 or a corresponding protein with a single-point mutation in the SH3 domain (pacsin-1P437L) was used. B. Liposome co-pelleting assay. Liposome binding assays were carried out with the complexes used in (A). The horizontal, dashed lines indicate the lipid-bound fraction of the isolated pacsin-1 F-BAR domain and isolated full-length pacsin-1 under similar conditions. Error bars represent standard deviations of a minimum of 3 independent experiments.
Figure 4.
The role of arginine residues within the dynamin-1 PRD in pacsin-1
's membrane deformation potential. A. Membrane deformation of Folch liposomes. The positions of Arg-to-Ala mutations (GST-PRDArgKO1, GST-PRDArgKO2 and GST-PRDArgKO3) in the mouse dynamin-1 PRD protein sequence are shown (top panel). Negative-stain EM images are shown after incubation of liposomes with the indicated protein complexes. B. Liposome co-pelleting assay. Liposome binding assays were carried out as described in Fig. 3. Error bars represent standard deviations of a minimum of 3 independent experiments.
Figure 5.
GST pull-down experiments were carried out by using wild-type and mutant forms of GST-PRD to examine their interactions with pacsin-1. Complexes were eluted and analyzed by SDS-PAGE and Coomassie-staining. Bait proteins: wild-type GST-PRD (PRDwt), GST-PRDtrunc1 (tr1), GST-PRDtrunc2 (tr2), GST-PRDArgKO1 (KO1), GST-PRDArgKO2 (KO2), GST-PRDArgKO3 (KO3) and GST (negative control).
Figure 6.
Activation of full-length endophilin-A1 by GST-PRD.
A. Negative-stain EM with Folch liposomes. Assays with endophilin-A1 (full-length or N-BAR domain; 10 µM) were carried out as described before. The inset shows a zoomed-in view of the red box area of the image, with scale bar = 100 nm. B. Statistical analysis of vesicle size distribution. Diameters of vesicles produced by endophilin in the presence of GST-PRD were quantified from electron micrographs taken from three independent experiments. C. Liposome co-pelleting assay with Folch liposomes. Liposome binding assays were carried out as described in Fig. 3. The horizontal, dashed lines indicate the lipid-bound fraction of the isolated endophilin-A1 N-BAR (expressed as His6-SUMO-fusion protein) domain and isolated full-length endophilin-A1 under similar conditions. Error bars represent standard deviations of a minimum of 3 independent experiments.
Figure 7.
Effect of liposome preparation method on protein-induced membrane deformation.
A. Negative-stain EM of liposomes prepared via sonication/ freeze-thaw (SFT), rapid solvent exchange (RSE), and freeze-thaw (FT) methods. B. Membrane deformation activities of various pacsin-1 constructs (5–10 µM) and CIP4F-BAR (10 µM) in Folch liposomes prepared following three different methods. Incubations and imaging were carried out as described previously.
Figure 8.
Effect of liposome diameter on protein-induced membrane deformation.
A. Negative-stain EM of extruded liposomes. Folch liposomes were prepared using the freeze-thaw (FT) method, followed by extrusion using pore sizes ranging from 100–1000 nm. Protein incubations and imaging was carried out as described above. B. Model of modulated, protein-induced membrane deformation potential. The schematic diagram illustrates the energies required to generate various membrane morphologies, which is likely dependent on the system's initial energy state. Considering only membrane properties and a constant number of lipid molecules in each system, more energy is needed to generate a defined number of smaller vesicles from larger, multi-lamellar liposomes, compared to smaller, uni-lamellar liposomes as the starting material. The system may also be subject to bimodality, where distinct structures (vesicle vs. tubule) could coexist as energetically equivalent structures.